Description of the model
September 30, 2007 - Updated specification for Model A, Michel Juillard model_A_sept07.pdf
September 10, 2007 - Updated specification for Model B, Wouter Den Hann modelb2007-09-10.pdf
May 28, 2007 - Michel Juillard's notes on problem A with suggestions for a smaller number of specifications: notes_mj.pdf. There may very well be still too many. Please comments according to your findings. Thanks. MJ. (Erratum in this document: in section 1.2, the second specification of utility function is wrong; should be 1+1/eta instead of 1-1/eta -- SébastienVillemot 2008-11-12 10:51:47)
- Document of April 2007 - Accuracy in models with heterogeneous agents and aggregate risk: Wouter Den Haan
Document of March 2007 - Problem B : http://www.cepremap.cnrs.fr/algan/testsuite/ModelB_2007.pdf
- Posted by Algan and Juillard (03/20/2007)
Document of July 2004: http://www.cepremap.cnrs.fr/juillard/testsuite/JEDC-ProbABC.pdf
Comments
- Document of April 2007 - Accuracy in models with heterogeneous agents and aggregate risk
Wouter Den Haan (06/04/2007)
This document shows that the traditional R-square measure might be misleading in assessing the accuracy in models with heterogeneous agents and aggregate risk. It provides an alternative procedure to check for accuracy that consists of alternative accuracy tests and exercises that can be used to confirm the tests or to evaluate the importance of possible inaccuracies.
- Document of March 2007: PROBLEM B - Heterogeneous agent models specification
Algan and Juillard (03/20/2007)
In SITE 2004, we agreed to proceed at least in three steps: - Solve the incomplete market economy with only idiosyncratic risks and exogenous interest rates. Compare the accuracy of the different methods to solve for individual policy rules. - Solve the incomplete market economy with only idiosyncratic risks but with endogenous interest rates. Proceed to the same accuracy comparison. - Solve the incomplete market economy with both aggregate risks and idiosyncratic risks. We agreed to all work with the same specification and to start from a common benchmark specification. The details of the model and of the parameterization are given in the attached document.